location portal's location-updated signal is never emited

[flatpak]

location portal's location-updated signal is never emited

[A6GibKm]

This is probably related to #37.

[grulja]

This should be fixed with #45. Isn't it?

[A6GibKm]

I tried to do this near the date the issue was opened, so probably not.

[grulja]

I tried to do this near the date the issue was opened, so probably not.

And do you use latest git or official releases? Because there was a release made after you opened this bug.

[A6GibKm]

Directly from master, unless GNOME Builder played some trick on me where it used an older cache (I don't think so, but it happens from time to time).

[A6GibKm]

I can confirm that it is not working on master.

[grulja]

Works for me.

I just added a simple code to Qt demo/test we have as part of libportal and I can see it sends the location-updated signal. Can you check this on DBus? I don't know Python so maybe you have a bug in your code? Also try to restart xdg-desktop-portal and start it with G_MESSAGES_DEBUG=all to see if there are some errors.

[A6GibKm]

I tried with vala, python, and the libportal demo (after adding a quick test), the code is executed and returns a success from the start_monitoring operation but the LocationUpdated signal is never emitted on the other side of the DBus.

At least on my laptop, it works fine on the first try of each login session, but then never again. On my desktop it fails randomly.

[grulja]

Try to use better precision and smaller threshold. Anyway, if you properly receive the signal at least one time, it means that the code in libportal does its job and there is no issue to be solved. If you have issues with location reporting then you probably need to report an issue in geoclue.

[A6GibKm]

Just for the record, unlike the test code above, all the local test are run with time_threshold = 0.

On the other hand ASHPD, a rust library to talk with portals, works all the time without any difficulty, so it is neither an issue in xdg-desktop-portals nor in geoclue.

[grulja]

Can you share output from dbus-monitor --session for both tests (ASHPD and libportal)?

[A6GibKm]

Here are the logs:

ashpd.txt
libportal.txt

The test was made with distance_threshold=60, time_threshold=0, and LocationAccuracy.CITY.

[A6GibKm]

Note that setting distance_threshold = 0 on libportal side will also make it work "reliably", but setting it to 1 will make it fail reliably on my machine after the first use on a session.

1. Reliable on my machine^tm.

[grulja]

I think I see the issue. I'll open a PR.

[grulja]

For the record, it's a bug in ASHPD and I already talked to @bilelmoussaoui about it. Basically there is a typo in both options so when you set distance_threshold and time_threshold it gets ignored and fallbacks to default (which is 0). Both options are actually distance-threshold and time-threshold → dash vs underscore.

Fixed with bilelmoussaoui/ashpd@817d1b7.